Open Data Market Reaches $350 Billion While Still Relying on Centralized Systems
The open data market has grown to an estimated $350 billion, but these resources largely depend on centralized infrastructure that contradicts Web3's philosophy of autonomy and censorship resistance. Michael O'Rourke, founder of Pocket Network and CEO of Grove, argues that open data must transition to decentralized infrastructure to fulfill its potential.
Decentralized infrastructure offers several advantages for open data applications. These include affordable large language model (LLM) training, accessible research data sharing, and reliable decentralized application (DApp) hosting. Each use case demonstrates why decentralized solutions provide more utility than centralized alternatives.
Open-source AI models like DeepSeek R1 cost approximately $5.5 million to train, compared to over $100 million for OpenAI's GPT-4. Despite these savings, the AI industry still depends on centralized platforms that conflict with open-source innovation. Decentralized networks can lower barriers by allowing operators to launch their own gateways.
Companies spend about $1 million daily on infrastructure to run LLM inference, creating a serviceable market of roughly $365 million annually. Platforms like Akash enable customized computing at 85% lower prices than centralized cloud providers, showing significant growth potential for decentralized solutions.
In scientific research, blockchain-based zero-knowledge machine learning models allow data sharing while preserving privacy. This approach enables researchers to access and share data without revealing sensitive information, bypassing expensive journal subscriptions that limit access to scientific knowledge.
Centralized hosting services have proven vulnerable to outages and access restrictions. In 2022, MetaMask users from certain regions lost access when Infura blocked them following US sanctions. Similar interruptions affected Infura clients in 2020, while Solana and Polygon experienced overloaded centralized remote procedure calls during high traffic periods. For Decentralized Autonomous Organizations (DAOs), these failures create operational risks since they rely on consistent network access for governance.
The Web3 RPC market processes approximately 100 billion serviceable requests, costing $3-$6 per million requests. This creates a total addressable market of $100-$200 million annually. With new data availability layers emerging, daily RPC requests could exceed 1 trillion. For DAOs specifically, decentralized infrastructure ensures their decision-making processes remain truly autonomous and resistant to external control or censorship.
Recent events have shown the practical consequences of centralization in supposedly decentralized projects. The TRON DAO Reserve shut down its governance portal for the USDD stablecoin in February 2025, despite marketing claims of being "community-driven." The portal was used only once before being removed, with major decisions including USDD 2.0 made without community input.
